INDEX

Explanations

phrases that indicate the existence or occurrence of events or situations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "was."

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

otic

-0.76

entric

-0.74

mare

-0.67

mania

-0.66

arta

-0.64

 selves

-0.63

asm

-0.62

 Yourself

-0.61

 Partner

-0.60

agos

-0.60

POSITIVE LOGITS

 murm

0.71

 plenty

0.70

wolves

0.69

TextColor

0.67

 GOODMAN

0.66

MpServer

0.65

owship

0.64

hes

0.64

 instrumental

0.63

ternity

0.63

Activations Density 0.061%

phrases that indicate the existence or occurrence of events or situations

instances of the word "was."

No Comments

No Known Activations