INDEX

Explanations

references to large-scale activities or entities

oai_token-act-pair · gpt-3.5-turbo

references to large-scale situations or phenomena

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

oother

-0.89

idges

-0.85

icut

-0.77

oras

-0.77

ĪĴ

-0.77

ston

-0.77

rium

-0.75

cair

-0.74

phis

-0.71

phe

-0.71

POSITIVE LOGITS

 fabrication

0.84

enter

0.80

scale

0.77

 scale

0.77

 production

0.75

 deployment

0.74

 fulfillment

0.74

 replica

0.73

 deployments

0.72

 administrative

0.69

Activations Density 0.020%

references to large-scale activities or entities

references to large-scale situations or phenomena

No Comments

No Known Activations