INDEX

Explanations

proper nouns, especially names of places and organizations

oai_token-act-pair · gpt-3.5-turbo

references to specific people or entities, particularly names or organizations

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

raq

-0.89

 Rockefeller

-0.87

Cro

-0.85

gra

-0.84

cro

-0.84

ran

-0.84

ctrl

-0.81

ra

-0.78

chu

-0.77

gow

-0.77

POSITIVE LOGITS

 hour

0.85

 Hours

0.82

Att

0.80

 Hour

0.80

 hours

0.79

 Archdemon

0.78

 alcohol

0.76

Element

0.75

 attire

0.73

 Ashton

0.73

Activations Density 0.341%

proper nouns, especially names of places and organizations

references to specific people or entities, particularly names or organizations

No Comments

No Known Activations