INDEX

Explanations

years, specifically looking for the year "2005"

oai_token-act-pair · gpt-3.5-turbo

references to the year 2005 and related data

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

oking

-0.75

sheet

-0.70

BOOK

-0.70

ebus

-0.68

Motion

-0.68

icz

-0.66

uve

-0.66

 Stru

-0.66

vere

-0.66

yang

-0.65

POSITIVE LOGITS

 ILCS

0.90

rak

0.90

-'

0.82

å¹

0.77

marked

0.73

acious

0.67

uated

0.67

 Lakes

0.65

rican

0.64

aciously

0.64

Activations Density 0.023%

years, specifically looking for the year "2005"

references to the year 2005 and related data

No Comments

No Known Activations