INDEX

Explanations

technical or structured information, possibly related to documents or reports

oai_token-act-pair · gpt-3.5-turbo

elements related to detailed descriptions and discussion prompts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

llan

-0.74

glers

-0.72

banks

-0.72

luck

-0.71

creation

-0.70

apesh

-0.69

dest

-0.68

finding

-0.67

spot

-0.66

Reply

-0.66

POSITIVE LOGITS

ctory

0.89

 pricing

0.85

 bundle

0.75

 Pricing

0.67

 discounts

0.66

 Thumbnails

0.66

 bundles

0.65

 aster

0.65

Ĥª

0.65

ãĥ¼ãĥ³

0.64

Activations Density 0.038%

technical or structured information, possibly related to documents or reports

elements related to detailed descriptions and discussion prompts

No Comments

No Known Activations