INDEX

Explanations

proper nouns, potentially related to companies, technologies, or organizations

oai_token-act-pair · gpt-3.5-turbo

letters and acronyms related to specific organizations or groups

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

taboola

-0.93

ItemTracker

-0.70

 showc

-0.68

ãĤ¨ãĥ«

-0.65

psc

-0.63

licted

-0.62

rounded

-0.62

enegger

-0.59

 compan

-0.59

fec

-0.59

POSITIVE LOGITS

senal

0.97

NT

0.96

Vs

0.96

AR

0.94

ATA

0.93

KA

0.91

TO

0.91

ENE

0.89

IA

0.88

OPS

0.87

Activations Density 0.129%

proper nouns, potentially related to companies, technologies, or organizations

letters and acronyms related to specific organizations or groups

No Comments

No Known Activations