INDEX

Explanations

numeric characters representing the value 10

oai_token-act-pair · gpt-3.5-turbo

distinct characters or symbols from various languages and scripts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ciating

-0.97

matically

-0.88

illac

-0.83

sterdam

-0.81

swick

-0.79

formance

-0.79

gdala

-0.77

versions

-0.76

brates

-0.75

anguage

-0.73

POSITIVE LOGITS

Î±

0.98

oti

0.92

Å«

0.77

Ð¾

0.77

Ð°

0.75

º

0.74

orter

0.73

·

0.72

uge

0.71

abba

0.71

Activations Density 0.007%

numeric characters representing the value 10

distinct characters or symbols from various languages and scripts

No Comments

No Known Activations