INDEX

Explanations

proper nouns, such as names of places, people, and organizations

oai_token-act-pair · gpt-3.5-turbo

terms related to inquiries and evaluations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 disabling

-0.62

 stimulating

-0.62

 Introduced

-0.60

CLASSIFIED

-0.57

 beginnings

-0.57

paren

-0.57

gradient

-0.57

 Allows

-0.56

 Casting

-0.55

ILA

-0.55

POSITIVE LOGITS

 belonged

0.95

 hail

0.89

 belong

0.85

 wore

0.84

 were

0.81

 consisted

0.80

owe

0.77

 behaved

0.77

 knew

0.75

 include

0.74

Activations Density 0.304%

proper nouns, such as names of places, people, and organizations

terms related to inquiries and evaluations

No Comments

No Known Activations