INDEX

Explanations

words or phrases related to bringing up or referencing particular topics or subjects

oai_token-act-pair · gpt-3.5-turbo

phrases that involve the action of raising a topic or issue for discussion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

jong

-0.66

=]

-0.65

 whisk

-0.64

 shipped

-0.64

Released

-0.60

BILITY

-0.59

arily

-0.59

 release

-0.59

arine

-0.58

 ASIC

-0.57

POSITIVE LOGITS

 amendments

0.71

osate

0.68

urb

0.66

ecided

0.66

ris

0.64

 grandchildren

0.63

TPP

0.63

 questions

0.62

ruciating

0.62

rir

0.61

Activations Density 0.021%

words or phrases related to bringing up or referencing particular topics or subjects

phrases that involve the action of raising a topic or issue for discussion

No Comments

No Known Activations