INDEX

Explanations

verbs in past tense

oai_token-act-pair · gpt-3.5-turbo

expressions of uncertainty or hesitation in opinions

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

 Gaddafi

-0.62

 Proposition

-0.59

 Originally

-0.56

 Vehicles

-0.56

moil

-0.55

raft

-0.55

 Manufact

-0.52

 Wales

-0.52

atories

-0.52

è£ħ

-0.52

POSITIVE LOGITS

 goddamn

0.89

 fucking

0.83

laughs

0.83

 fuckin

0.81

 anyways

0.80

 damn

0.79

 darn

0.78

 godd

0.74

 haha

0.73

.).

0.72

Activations Density 1.968%

verbs in past tense

expressions of uncertainty or hesitation in opinions

No Comments

No Known Activations