INDEX

Explanations

words related to actions or states involving secret or hidden intentions

oai_token-act-pair · gpt-3.5-turbo

verbs and gerunds that indicate ongoing actions or processes

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

reported

-0.61

 Lauder

-0.60

uez

-0.59

WTC

-0.58

ãĤ¨ãĥ«

-0.58

////////////////

-0.58

REP

-0.58

Ry

-0.58

ngth

-0.57

OY

-0.56

POSITIVE LOGITS

levard

0.97

ĪĴ

0.82

ĵĺ

0.82

lehem

0.81

aneers

0.80

apest

0.80

halla

0.76

ategory

0.70

pillar

0.70

hammad

0.70

Activations Density 0.073%

words related to actions or states involving secret or hidden intentions

verbs and gerunds that indicate ongoing actions or processes

No Comments

No Known Activations