INDEX

Explanations

phrases related to legal and judicial contexts

oai_token-act-pair · gpt-3.5-turbo

instances of legal terminology related to immunity and liability

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

.<

-0.57

'.

-0.55

 instead

-0.53

maxwell

-0.53

terness

-0.52

".

-0.52

''.

-0.52

$.

-0.52

.).

-0.52

ucci

-0.52

POSITIVE LOGITS

pires

0.80

mentioned

0.55

 nutshell

0.54

 depends

0.52

 differed

0.50

 varies

0.49

 comprises

0.48

pired

0.48

vernment

0.47

 relates

0.45

Activations Density 1.463%

phrases related to legal and judicial contexts

instances of legal terminology related to immunity and liability

No Comments

No Known Activations