INDEX

Explanations

mentions of people or entities delivering official statements or letters to others

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "to."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 deductions

-0.74

ahime

-0.71

illac

-0.71

 wastes

-0.69

ain

-0.67

lux

-0.66

 bets

-0.66

 folds

-0.66

tions

-0.64

 prizes

-0.62

POSITIVE LOGITS

 accompany

1.22

 commemorate

1.05

 appease

0.92

 coincide

0.88

 signify

0.88

wered

0.88

 clarify

0.87

 reassure

0.86

 celebrate

0.86

 satisfy

0.84

Activations Density 0.152%

mentions of people or entities delivering official statements or letters to others

occurrences of the word "to."

No Comments

No Known Activations