INDEX

Explanations

numerical data such as statistics

oai_token-act-pair · gpt-3.5-turbo

statistical references and mentions of eavesdropping

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

ality

-0.86

esan

-0.80

utral

-0.77

hyde

-0.74

ciating

-0.72

icate

-0.72

ism

-0.72

icity

-0.70

ships

-0.69

mental

-0.68

POSITIVE LOGITS

hib

0.76

rador

0.76

ENSE

0.73

--+

0.72

ORTS

0.72

ENDED

0.71

ordinary

0.70

ouls

0.69

oul

0.69

ãĤº

0.68

Activations Density 0.048%

numerical data such as statistics

statistical references and mentions of eavesdropping

No Comments

No Known Activations