INDEX

Explanations

phrases related to insider information or knowledge about specific situations or matters

oai_token-act-pair · gpt-3.5-turbo

phrases related to sources of information and knowledge in discussions or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĤª

-0.67

Lv

-0.63

Bec

-0.58

Quantity

-0.57

];

-0.57

 Siber

-0.56

Purchase

-0.55

!).

-0.54

Chart

-0.54

obal

-0.54

POSITIVE LOGITS

 told

1.19

 cautioned

1.13

 said

1.11

say

1.08

 remarked

1.04

 thanked

1.02

 commented

1.01

 testified

1.00

 urged

0.99

told

0.98

Activations Density 0.210%

phrases related to insider information or knowledge about specific situations or matters

phrases related to sources of information and knowledge in discussions or events

No Comments

No Known Activations