INDEX

Explanations

phrases or sentences discussing trends or patterns in data

oai_token-act-pair · gpt-3.5-turbo

recurring mentions of trends

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gur

-0.69

vette

-0.67

 Purchase

-0.67

dos

-0.65

akov

-0.63

 oath

-0.62

Lak

-0.62

ument

-0.61

 expend

-0.60

ighth

-0.59

POSITIVE LOGITS

setting

1.30

 toward

1.19

 towards

1.09

 trend

1.01

chart

0.88

lines

0.88

set

0.88

iness

0.87

 downwards

0.85

 reversal

0.85

Activations Density 0.054%

phrases or sentences discussing trends or patterns in data

recurring mentions of trends

No Comments

No Known Activations