INDEX

Explanations

numbers indicating points being scored in different contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "score."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

agan

-0.81

etheless

-0.61

 hind

-0.61

lly

-0.60

perial

-0.58

 Society

-0.57

IDA

-0.57

conn

-0.57

por

-0.57

 pleas

-0.56

POSITIVE LOGITS

 score

1.21

 scores

1.08

 Score

1.01

ificant

0.93

card

0.91

keeper

0.88

 Scores

0.87

 scored

0.83

cards

0.80

Score

0.80

Activations Density 0.010%

numbers indicating points being scored in different contexts

instances of the word "score."

No Comments

No Known Activations