INDEX

Explanations

numbers that appear to be part of specific coded messages or data

oai_token-act-pair · gpt-3.5-turbo

specific numerical identifiers or codes

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

ging

-0.81

 snipp

-0.77

leck

-0.74

ged

-0.72

nington

-0.72

anke

-0.70

gotten

-0.70

uyomi

-0.69

inition

-0.69

akin

-0.68

POSITIVE LOGITS

 ILCS

1.08

mph

0.99

dB

0.88

dB

0.85

rup

0.85

0.79

rollers

0.76

0.73

0.72

Activations Density 0.026%

numbers that appear to be part of specific coded messages or data

specific numerical identifiers or codes

No Comments

No Known Activations