INDEX

Explanations

mentions of layers or levels with high emphasis or importance

oai_token-act-pair · gpt-3.5-turbo

references to layers in various contexts

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

ãĥīãĥ©

-0.84

 Predators

-0.75

speaking

-0.74

 STAR

-0.72

orie

-0.71

date

-0.70

lishing

-0.70

FRE

-0.66

itarian

-0.66

Rao

-0.65

POSITIVE LOGITS

 layers

1.35

 layer

1.21

 thickness

0.99

layer

0.91

Layer

0.87

 Layer

0.83

 opacity

0.80

coat

0.79

 coats

0.77

 veil

0.77

Activations Density 0.019%

mentions of layers or levels with high emphasis or importance

references to layers in various contexts

No Comments

No Known Activations