INDEX

Explanations

instances where someone or something is described as prepared or willing

oai_token-act-pair · gpt-3.5-turbo

instances of the word "ready."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

regate

-0.78

oret

-0.75

roma

-0.73

apple

-0.68

uo

-0.66

orb

-0.64

efficiency

-0.62

omal

-0.61

 sting

-0.60

effects

-0.60

POSITIVE LOGITS

Ready

0.81

 Ready

0.77

 SERV

0.76

 ready

0.76

ÃįÃį

0.69

 mentally

0.69

arity

0.67

 Dragonbound

0.66

 supper

0.66

ptin

0.63

Activations Density 0.016%

instances where someone or something is described as prepared or willing

instances of the word "ready."

No Comments

No Known Activations