INDEX

Explanations

phrases indicating strong recommendations or pleas for action

oai_token-act-pair · gpt-3.5-turbo

instances of the word "urge" and its variations in the context of requesting or recommending actions

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

çĦ

-0.72

ìĿ

-0.68

sf

-0.63

anuts

-0.63

central

-0.62

ëĭ

-0.61

 Kear

-0.61

bourne

-0.61

Anderson

-0.59

miah

-0.59

POSITIVE LOGITS

 caution

0.93

FontSize

0.78

 forgiveness

0.74

 vigilance

0.72

 urges

0.72

 urge

0.72

 patience

0.71

 anew

0.71

 pressuring

0.70

pread

0.70

Activations Density 0.035%

phrases indicating strong recommendations or pleas for action

instances of the word "urge" and its variations in the context of requesting or recommending actions

No Comments

No Known Activations