INDEX

Explanations

terms related to different categories such as race, class, gender, and other characteristics within a social context

oai_token-act-pair · gpt-3.5-turbo

references to social categorizations and roles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

urtles

-0.76

DonaldTrump

-0.75

Newsletter

-0.67

bledon

-0.63

 ÂŃ

-0.61

 Remem

-0.59

Salt

-0.59

isSpecialOrderable

-0.56

Send

-0.55

displayText

-0.55

POSITIVE LOGITS

/,

1.73

1.68

/)

1.55

/"

1.53

/?

1.49

/_

1.46

/(

1.43

/#

1.42

/.

1.40

 combo

1.32

Activations Density 0.130%

terms related to different categories such as race, class, gender, and other characteristics within a social context

references to social categorizations and roles

No Comments

No Known Activations