INDEX

Explanations

references to different forms of something

oai_token-act-pair · gpt-3.5-turbo

phrases that describe different types or categories of something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Kingdoms

-0.87

iets

-0.84

esses

-0.81

heads

-0.81

wives

-0.80

naires

-0.79

stones

-0.79

lists

-0.76

parts

-0.76

estones

-0.75

POSITIVE LOGITS

 discrimination

0.96

 oppression

0.95

 transportation

0.94

 harassment

0.89

 coercion

0.89

 persuasion

0.89

 abuse

0.88

 communication

0.87

 entertainment

0.87

 imprisonment

0.87

Activations Density 0.071%

references to different forms of something

phrases that describe different types or categories of something

No Comments

No Known Activations