INDEX

Explanations

instances of attributing statements or actions to specific sources

oai_token-act-pair · gpt-3.5-turbo

phrases that reference sources or authorities, particularly using the phrase "according to."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ograp

-0.78

obyl

-0.76

 cared

-0.71

gage

-0.71

anooga

-0.67

icably

-0.67

oor

-0.66

itaire

-0.66

ocr

-0.64

 accommod

-0.64

POSITIVE LOGITS

 sources

0.97

 STATS

0.80

 Reuters

0.78

 Variety

0.78

 insiders

0.74

 reports

0.73

reports

0.72

 Folk

0.72

source

0.72

 Kant

0.72

Activations Density 0.065%

instances of attributing statements or actions to specific sources

phrases that reference sources or authorities, particularly using the phrase "according to."

No Comments

No Known Activations