INDEX

Explanations

instances where something is expected to happen or be a certain way

oai_token-act-pair · gpt-3.5-turbo

phrases indicating predictions or expectations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Sure

-0.72

amps

-0.66

FX

-0.64

 Connection

-0.62

checking

-0.62

Choice

-0.61

 imagining

-0.61

 Checking

-0.61

 Promotion

-0.60

-0.59

POSITIVE LOGITS

be

1.13

 continue

1.08

 arrive

1.04

 settle

1.02

 prevail

1.00

 announce

1.00

 accelerate

0.97

 reap

0.96

 comply

0.95

 intensify

0.95

Activations Density 0.106%

instances where something is expected to happen or be a certain way

phrases indicating predictions or expectations

No Comments

No Known Activations