INDEX

Explanations

terms related to classification or grouping based on certain characteristics or features

oai_token-act-pair · gpt-3.5-turbo

mentions of different types or categories

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

 Prospect

-0.64

SPONSORED

-0.64

bats

-0.63

 Berger

-0.62

 Boxing

-0.62

other

-0.61

 pollen

-0.59

Beast

-0.59

erb

-0.58

une

-0.57

POSITIVE LOGITS

 halves

0.97

 sexes

0.89

 thirds

0.89

 totaling

0.85

rencies

0.80

 simultaneously

0.79

 separated

0.79

 apiece

0.78

agos

0.75

 sets

0.73

Activations Density 0.346%

terms related to classification or grouping based on certain characteristics or features

mentions of different types or categories

No Comments

No Known Activations