INDEX

Explanations

personal pronouns and verbs related to reasons or explanations

oai_token-act-pair · gpt-3.5-turbo

pronouns and references to individuals in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 sqor

-0.76

dds

-0.72

ĺħ

-0.68

 ILCS

-0.66

oret

-0.63

 courtesy

-0.62

FORMATION

-0.61

OSS

-0.61

¶ħ

-0.60

eatures

-0.59

POSITIVE LOGITS

 chose

1.23

 decided

0.99

 bother

0.96

 bothered

0.93

 opted

0.91

 mattered

0.90

 Matters

0.88

 couldn

0.88

 hesitated

0.87

 bothers

0.85

Activations Density 0.141%

personal pronouns and verbs related to reasons or explanations

pronouns and references to individuals in the text

No Comments

No Known Activations