INDEX

Explanations

references to significant or notable aspects or achievements

oai_token-act-pair · gpt-3.5-turbo

phrases highlighting superlatives and the most significant attributes or achievements

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

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

arth

-0.82

 Bundy

-0.79

 Canaver

-0.77

skirts

-0.76

abus

-0.74

arthy

-0.70

ards

-0.69

witz

-0.69

uador

-0.68

 Guan

-0.66

POSITIVE LOGITS

 adversary

0.95

 adversaries

0.89

 selves

0.82

 counterparts

0.81

 foray

0.78

iaries

0.77

 predecessors

0.76

 colleague

0.72

 counterpart

0.72

 confid

0.71

Activations Density 0.104%

references to significant or notable aspects or achievements

phrases highlighting superlatives and the most significant attributes or achievements

No Comments

No Known Activations