INDEX

Explanations

references to different types of entities, emotions, or concepts

oai_token-act-pair · gpt-3.5-turbo

terms associated with animals and humans in various contexts

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

 GOODMAN

-0.79

Effective

-0.78

ificantly

-0.73

orthy

-0.71

oubted

-0.66

pecially

-0.66

cially

-0.64

 concess

-0.63

relevant

-0.62

 Specific

-0.62

POSITIVE LOGITS

stros

0.86

 awaits

0.82

smanship

0.81

kid

0.76

oids

0.74

scape

0.73

kick

0.70

pour

0.68

 sandwic

0.67

ulence

0.66

Activations Density 0.704%

references to different types of entities, emotions, or concepts

terms associated with animals and humans in various contexts

No Comments

No Known Activations