INDEX

Explanations

words relating to removing or avoiding obstacles or constraints

oai_token-act-pair · gpt-3.5-turbo

references to de-anonymization techniques

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

vernment

-0.70

xxx

-0.69

estial

-0.68

Fry

-0.68

tyard

-0.68

xx

-0.67

zag

-0.67

RD

-0.64

cial

-0.64

housing

-0.63

POSITIVE LOGITS

ãĤ£

0.99

afia

0.86

ovember

0.85

aintain

0.84

ploy

0.84

antle

0.83

ikhail

0.82

 Nadu

0.81

iami

0.77

asking

0.77

Activations Density 0.034%

words relating to removing or avoiding obstacles or constraints

references to de-anonymization techniques

No Comments

No Known Activations