INDEX

Explanations

verbs related to actions of clearing or removing obstacles or objects

oai_token-act-pair · gpt-3.5-turbo

instances of the word "clear"

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

eatures

-0.69

ITAL

-0.68

Bey

-0.67

ENTS

-0.66

 nostalg

-0.66

 Loft

-0.63

ENCE

-0.63

 inval

-0.61

QL

-0.60

Lenin

-0.60

POSITIVE LOGITS

ances

1.18

ance

0.94

cuts

0.92

houses

0.91

cloth

0.89

 clears

0.83

cutting

0.79

cut

0.78

boards

0.77

brush

0.77

Activations Density 0.022%

verbs related to actions of clearing or removing obstacles or objects

instances of the word "clear"

No Comments

No Known Activations