INDEX

Explanations

phrases emphasizing contrast or negation

oai_token-act-pair · gpt-3.5-turbo

negations and expressions of denial

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

psey

-0.64

ce

-0.62

often

-0.61

typically

-0.59

Ens

-0.56

 Advertisement

-0.56

 airs

-0.56

 shaman

-0.56

peace

-0.56

psi

-0.56

POSITIVE LOGITS

onen

0.76

\\\\\\\\

0.69

ounters

0.68

ibe

0.67

 benefited

0.66

omical

0.66

gged

0.65

rist

0.64

å¤

0.63

idon

0.63

Activations Density 0.262%

phrases emphasizing contrast or negation

negations and expressions of denial

No Comments

No Known Activations