INDEX

Explanations

words expressing strong emotions or preferences towards certain actions or entities

oai_token-act-pair · gpt-3.5-turbo

expressions of affection and aversion

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

phabet

-0.81

EStreamFrame

-0.77

rome

-0.75

externalActionCode

-0.72

harm

-0.69

soType

-0.66

nw

-0.65

level

-0.65

levels

-0.65

aqu

-0.64

POSITIVE LOGITS

 themselves

0.67

 Pigs

0.61

 revenge

0.59

 foreigners

0.59

 passionately

0.59

 sticking

0.58

 eagerly

0.58

 outsiders

0.58

 importing

0.56

 advertising

0.56

Activations Density 0.319%

words expressing strong emotions or preferences towards certain actions or entities

expressions of affection and aversion

No Comments

No Known Activations