INDEX

Explanations

phrases related to the style or method in which actions are performed

oai_token-act-pair · gpt-3.5-turbo

phrases that describe the way actions are performed

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

Dust

-0.77

rament

-0.74

Lyn

-0.71

ILL

-0.69

inka

-0.68

minster

-0.67

hemat

-0.66

 Patri

-0.66

tek

-0.65

ouston

-0.65

POSITIVE LOGITS

isms

1.09

 manner

0.85

 dictated

0.83

 largeDownload

0.78

abus

0.77

able

0.76

ality

0.76

 fashion

0.74

ably

0.73

ebin

0.73

Activations Density 0.017%

phrases related to the style or method in which actions are performed

phrases that describe the way actions are performed

No Comments

No Known Activations