INDEX

Explanations

words related to problematic situations or conflicts

oai_token-act-pair · gpt-3.5-turbo

references to trouble and its potential consequences

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

oliberal

-0.88

ulture

-0.75

uine

-0.74

hetics

-0.72

ithe

-0.70

ivist

-0.68

orney

-0.68

entric

-0.67

audi

-0.67

ITAL

-0.67

POSITIVE LOGITS

hooting

1.05

 troubles

0.90

makers

0.85

 trouble

0.81

maker

0.77

fully

0.75

some

0.73

 interfering

0.70

flake

0.69

making

0.69

Activations Density 0.018%

words related to problematic situations or conflicts

references to trouble and its potential consequences

No Comments

No Known Activations