INDEX

Explanations

references to various "Mission" entities or projects

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

locations or places

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

manship

-0.75

 doping

-0.68

 jaws

-0.66

sol

-0.64

 antitrust

-0.62

 COVER

-0.61

 electronically

-0.61

rums

-0.60

lection

-0.60

soever

-0.60

POSITIVE LOGITS

aries

1.20

ary

1.02

arily

0.86

acular

0.82

atorium

0.82

 Impossible

0.80

aires

0.79

 Mission

0.79

adelphia

0.79

Vie

0.78

Activations Density 0.037%

references to various "Mission" entities or projects

locations or places

No Comments

No Known Activations