INDEX

Explanations

phrases or sentences that mention a wide range of items or topics

oai_token-act-pair · gpt-3.5-turbo

discussions about a wide variety of topics

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

wards

-0.73

IDA

-0.73

cel

-0.72

si

-0.72

scape

-0.70

bed

-0.69

yl

-0.67

rafted

-0.67

bal

-0.67

mit

-0.66

POSITIVE LOGITS

 ranging

1.15

 ranges

0.94

 ranged

0.88

 range

0.77

ãĤ¤ãĥĪ

0.76

 spanning

0.72

isode

0.72

ranging

0.71

 fortun

0.71

 conduc

0.71

Activations Density 0.009%

phrases or sentences that mention a wide range of items or topics

discussions about a wide variety of topics

No Comments

No Known Activations