INDEX

Explanations

word sequences related to layers in a technical or metaphorical sense

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 @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

orie

-0.76

ãĥīãĥ©

-0.75

 Springer

-0.73

date

-0.72

eln

-0.72

ivia

-0.72

oris

-0.70

ILLE

-0.70

iggurat

-0.69

 Reeves

-0.67

POSITIVE LOGITS

 thickness

1.20

 layers

1.20

 layer

1.07

coat

0.85

layer

0.84

arity

0.83

 cake

0.82

 opacity

0.80

 layered

0.78

 underneath

0.77

Activations Density 0.062%

word sequences related to layers in a technical or metaphorical sense

references to layers in various contexts

No Comments

No Known Activations