INDEX

Explanations

numbers or symbols that are not typical in natural language text

oai_token-act-pair · gpt-3.5-turbo

sections of text that include high-frequency or common phrases and possibly denote general statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 tack

-0.83

 estab

-0.75

 exha

-0.73

 interven

-0.73

 prec

-0.72

 antit

-0.71

 tram

-0.71

 sculpt

-0.71

 stabil

-0.69

 overpower

-0.69

POSITIVE LOGITS

When

1.40

Sometimes

1.40

Yesterday

1.35

Whether

1.34

Today

1.33

If

1.33

Unless

1.31

Whenever

1.30

Following

1.30

There

1.29

Activations Density 0.208%

numbers or symbols that are not typical in natural language text

sections of text that include high-frequency or common phrases and possibly denote general statements

No Comments

No Known Activations