INDEX

Explanations

numerical values and abbreviations related to financial institutions or processes

oai_token-act-pair · gpt-3.5-turbo

references to specific datasets or systems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ãĤ©

-0.94

 thous

-0.75

ãĥĪ

-0.68

ãĥł

-0.66

combe

-0.65

ãĤ¹

-0.61

ãĥī

-0.61

aved

-0.59

ãĥĸ

-0.57

taboola

-0.57

POSITIVE LOGITS

atcher

0.82

HS

0.82

enges

0.77

dit

0.77

OPLE

0.75

ZI

0.74

UTE

0.72

JD

0.69

 Reloaded

0.69

OLOGY

0.68

Activations Density 0.267%

numerical values and abbreviations related to financial institutions or processes

references to specific datasets or systems

No Comments

No Known Activations