INDEX

Explanations

personal pronouns and possessive pronouns referring to someone else

oai_token-act-pair · gpt-3.5-turbo

statements related to blame and accountability

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

hess

-0.56

 Milton

-0.53

cas

-0.52

NEC

-0.50

Tut

-0.50

 Klaus

-0.48

 Hammond

-0.48

 Prix

-0.47

 Formula

-0.47

 Gates

-0.46

POSITIVE LOGITS

*/(

0.76

EStream

0.71

ymes

0.69

É

0.66

awaru

0.65

laughs

0.64

ngth

0.63

âĹ¼

0.63

ãĥ³ãĤ¸

0.62

INO

0.61

Activations Density 2.600%

personal pronouns and possessive pronouns referring to someone else

statements related to blame and accountability

No Comments

No Known Activations