INDEX

Explanations

adjectives related to "unknown" or "not yet completed"

oai_token-act-pair · gpt-3.5-turbo

terms related to unresolved issues or injustices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ensing

-0.87

cium

-0.86

cius

-0.85

esthetic

-0.85

aptic

-0.84

reens

-0.84

ldon

-0.83

phas

-0.81

ingers

-0.81

etics

-0.81

POSITIVE LOGITS

 untold

0.86

 unheard

0.76

 srfAttach

0.73

 Crimes

0.69

 Johann

0.69

 Cinderella

0.68

ILLE

0.68

accompan

0.67

CLASSIFIED

0.66

 Chero

0.66

Activations Density 0.017%

adjectives related to "unknown" or "not yet completed"

terms related to unresolved issues or injustices

No Comments

No Known Activations