INDEX

Explanations

unique identifiers or specific details within a larger context

oai_token-act-pair · gpt-3.5-turbo

references to explanatory clauses beginning with "which."

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

Tes

-0.73

Behind

-0.69

GY

-0.69

greg

-0.68

CLOSE

-0.65

Uk

-0.65

athi

-0.65

 Hatt

-0.64

rior

-0.64

prototype

-0.64

POSITIVE LOGITS

soever

0.88

 derives

0.72

 incidentally

0.72

xual

0.71

wikipedia

0.68

 dearly

0.68

 consists

0.67

akespeare

0.67

 admittedly

0.66

 consisted

0.65

Activations Density 0.027%

unique identifiers or specific details within a larger context

references to explanatory clauses beginning with "which."

No Comments

No Known Activations