INDEX

Explanations

programming or technical commands enclosed in symbols like '>' and numbers

oai_token-act-pair · gpt-3.5-turbo

section headers or navigational cues in instructional or technical documents

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

 destro

-0.95

ModLoader

-0.93

onial

-0.90

etheless

-0.89

erville

-0.85

teenth

-0.85

yssey

-0.84

iqueness

-0.83

acies

-0.81

hement

-0.81

POSITIVE LOGITS

_>

1.22

=>

0.84

++++

0.82

0.76

.<

0.75

 Dangerous

0.69

 Preferences

0.66

tf

0.66

(<

0.65

>>>>

0.65

Activations Density 0.014%

programming or technical commands enclosed in symbols like '>' and numbers

section headers or navigational cues in instructional or technical documents

No Comments

No Known Activations