INDEX

Explanations

phrases related to news articles or reports

oai_token-act-pair · gpt-3.5-turbo

high-frequency numerical indicators of significance or support

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

Ĥ

-0.85

poon

-0.78

İ

-0.75

©

-0.74

 redes

-0.73

ettings

-0.72

¼

-0.71

¥µ

-0.71

Ĥ¬

-0.69

ĳ

-0.68

POSITIVE LOGITS

FILE

1.24

ÂłÂłÂłÂłÂłÂłÂłÂł

0.80

Statement

0.80

Credit

0.79

Returns

0.76

IJ

0.76

Cond

0.74

Fresh

0.73

However

0.73

Despite

0.73

Activations Density 0.072%

phrases related to news articles or reports

high-frequency numerical indicators of significance or support

No Comments

No Known Activations