INDEX

Explanations

phrases where an action or statement is met with a specific response or reaction

oai_token-act-pair · gpt-3.5-turbo

phrases indicating reception or response to events or statements

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

 Pastebin

-0.70

 prefrontal

-0.68

 republic

-0.65

XV

-0.64

 principals

-0.64

 situational

-0.63

 dosage

-0.63

 jurors

-0.63

 Paste

-0.63

 vantage

-0.62

POSITIVE LOGITS

idable

0.96

ãĤ¦ãĤ¹

0.84

alled

0.82

å£

0.82

oing

0.82

ª

0.82

igated

0.80

gaard

0.78

DragonMagazine

0.78

ãĥīãĥ©

0.78

Activations Density 0.101%

phrases where an action or statement is met with a specific response or reaction

phrases indicating reception or response to events or statements

No Comments

No Known Activations