INDEX

Explanations

programming syntax and keywords

oai_token-act-pair · gpt-3.5-turbo

symbols or formatting elements in text, specifically identifying undefined states or errors in programming context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 acknowled

-0.75

 imaginary

-0.73

 Guarant

-0.68

 Â©

-0.68

 estab

-0.66

hay

-0.63

 guarantee

-0.61

Ãī

-0.61

 princ

-0.60

tv

-0.59

POSITIVE LOGITS

bably

0.79

furt

0.77

atana

0.75

esson

0.75

bats

0.74

livious

0.73

Lex

0.73

esides

0.73

SPA

0.73

 Samoa

0.72

Activations Density 0.000%

programming syntax and keywords

symbols or formatting elements in text, specifically identifying undefined states or errors in programming context

No Comments

No Known Activations