INDEX

Explanations

words related to legal language or implications

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "guilty"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

cling

-0.77

croft

-0.75

 Spectrum

-0.73

ŃĶ

-0.73

 spect

-0.67

riad

-0.66

hower

-0.65

HAEL

-0.65

cycle

-0.62

rums

-0.61

POSITIVE LOGITS

idelines

1.12

pta

1.09

errilla

1.07

vernment

1.07

arding

1.07

ilty

1.05

arant

1.03

cci

1.01

inea

0.98

bernatorial

0.95

Activations Density 0.015%

words related to legal language or implications

occurrences of the word "guilty"

No Comments

No Known Activations