INDEX

Explanations

numerical values preceded by special characters that have been stylized

oai_token-act-pair · gpt-3.5-turbo

instances of numerical data or statistics

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

 Kinder

-0.73

 Blumenthal

-0.69

 Able

-0.65

 herself

-0.59

 Sind

-0.59

 Niet

-0.59

 grow

-0.59

 Karachi

-0.58

Nep

-0.58

 Mata

-0.58

POSITIVE LOGITS

Anyway

0.90

Reason

0.79

morrow

0.77

 Anyway

0.76

[[

0.76

================================================================

0.74

ï¸ı

0.72

ï¸

0.72

tree

0.71

{{

0.71

Activations Density 0.355%

numerical values preceded by special characters that have been stylized

instances of numerical data or statistics

No Comments

No Known Activations