INDEX

Explanations

terms related to historical events and organizations, especially those with specific abbreviations or acronyms

oai_token-act-pair · gpt-3.5-turbo

references to geographic or organizational designations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Franch

-0.73

idates

-0.72

uate

-0.71

wagen

-0.71

Nig

-0.70

 fitting

-0.70

utive

-0.69

uated

-0.67

holders

-0.66

ulator

-0.66

POSITIVE LOGITS

WW

1.23

II

1.13

III

1.13

JD

1.09

WF

1.05

LD

1.04

MJ

1.03

BF

0.99

OA

0.99

TW

0.97

Activations Density 0.020%

terms related to historical events and organizations, especially those with specific abbreviations or acronyms

references to geographic or organizational designations

No Comments

No Known Activations