INDEX

Explanations

technical or procedural terms related to processes

oai_token-act-pair · gpt-3.5-turbo

mentions of processes and procedures

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

hiro

-0.68

azo

-0.66

cknowled

-0.65

 dearly

-0.64

ippers

-0.64

ged

-0.62

gdala

-0.62

legged

-0.61

 railing

-0.61

 Haram

-0.60

POSITIVE LOGITS

ional

1.09

ions

0.85

ivity

0.84

amins

0.82

ioned

0.78

 Process

0.77

ebin

0.76

ivism

0.76

APH

0.75

 processes

0.74

Activations Density 0.026%

technical or procedural terms related to processes

mentions of processes and procedures

No Comments

No Known Activations