INDEX

Explanations

terms related to "factor", especially emphasizing on numbers or quantification

oai_token-act-pair · gpt-3.5-turbo

references to factors, often in a numerical or comparative context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

escription

-0.86

gem

-0.86

 Seym

-0.79

alach

-0.79

quartered

-0.78

ccording

-0.78

 Parables

-0.76

gements

-0.76

sterdam

-0.75

pit

-0.72

POSITIVE LOGITS

 factor

0.99

factor

0.97

 factors

0.87

Factor

0.77

inity

0.77

 Factor

0.75

 equation

0.75

oxin

0.72

 Factors

0.71

iator

0.69

Activations Density 0.010%

terms related to "factor", especially emphasizing on numbers or quantification

references to factors, often in a numerical or comparative context

No Comments

No Known Activations