INDEX

Explanations

instances where a decision has been made between different options

oai_token-act-pair · gpt-3.5-turbo

instances of the word "opted" along with its variations, indicating choices or decisions made

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

counter

-0.71

strength

-0.71

dist

-0.67

 crush

-0.66

mean

-0.64

 literature

-0.62

 Defeat

-0.62

iot

-0.61

ance

-0.61

 extingu

-0.60

POSITIVE LOGITS

 opted

3.20

 opting

2.53

 resorted

1.11

 preferring

1.11

 succumbed

1.11

 esche

1.09

 transitioned

1.04

opt

1.02

 teamed

1.00

 swapped

0.99

Activations Density 0.027%

instances where a decision has been made between different options

instances of the word "opted" along with its variations, indicating choices or decisions made

No Comments

No Known Activations

instances where a decision has been made between different options

instances of the word "opted" along with its variations, indicating choices or decisions made

No Comments

No Known Activations