INDEX

Explanations

instances referencing a step-by-step procedure or progression

oai_token-act-pair · gpt-3.5-turbo

mentions of processes or procedural terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

akening

-0.71

ighter

-0.69

collar

-0.65

IGHTS

-0.65

 cheap

-0.62

 disposable

-0.61

 hearty

-0.61

 trad

-0.61

 complimentary

-0.61

 earnest

-0.60

POSITIVE LOGITS

 process

1.01

ional

0.94

 processes

0.92

 Process

0.90

ioned

0.87

usercontent

0.85

naires

0.83

process

0.82

 Procedure

0.80

ions

0.80

Activations Density 0.027%

instances referencing a step-by-step procedure or progression

mentions of processes or procedural terms

No Comments

No Known Activations