INDEX

Explanations

keywords related to technical details, such as dates, figures, and instructions

oai_token-act-pair · gpt-3.5-turbo

instances of significant events or occurrences, particularly related to health and law

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

'."

-0.63

 mould

-0.54

 colour

-0.49

 flavours

-0.48

 flavour

-0.44

 hazard

-0.44

 civilisation

-0.44

 learnt

-0.43

',"

-0.43

 offences

-0.42

POSITIVE LOGITS

 Â·

0.49

ometime

0.47

 livest

0.45

âģ

0.42

\.

0.41

endo

0.40

alone

0.40

 Starts

0.40

 âĢķ

0.40

bernatorial

0.40

Activations Density 2.691%

keywords related to technical details, such as dates, figures, and instructions

instances of significant events or occurrences, particularly related to health and law

No Comments

No Known Activations