INDEX

Explanations

phrases related to alcohol intoxication

oai_token-act-pair · gpt-3.5-turbo

references to drunkenness and related behaviors or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

atown

-0.87

acea

-0.85

aeda

-0.83

achev

-0.82

aceous

-0.80

ilege

-0.79

chio

-0.79

ARI

-0.75

isp

-0.74

orescent

-0.73

POSITIVE LOGITS

ness

0.95

ly

0.92

liness

0.86

furt

0.77

scribe

0.70

 drunken

0.70

st

0.67

ths

0.66

shire

0.66

stery

0.65

Activations Density 0.026%

phrases related to alcohol intoxication

references to drunkenness and related behaviors or situations

No Comments

No Known Activations