INDEX

Explanations

words related to text formatting or special characters

oai_token-act-pair · gpt-3.5-turbo

instances of the letter "L" in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 looms

-0.64

 strat

-0.63

 paralle

-0.63

 Tulsa

-0.59

 spir

-0.59

 Negro

-0.59

 paved

-0.58

 pres

-0.58

 controvers

-0.58

 collapses

-0.57

POSITIVE LOGITS

ï¸ı

1.34

who

0.88

rative

0.85

ship

0.84

selves

0.83

ski

0.82

iversary

0.79

resent

0.78

ufact

0.77

tre

0.76

Activations Density 0.261%

words related to text formatting or special characters

instances of the letter "L" in various contexts

No Comments

No Known Activations