INDEX

Explanations

pronouns referring to the self or specific individuals

oai_token-act-pair · gpt-3.5-turbo

pronouns indicating the speaker or listener

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

jri

-0.81

aukee

-0.74

ibaba

-0.73

downs

-0.67

dden

-0.66

acular

-0.66

icist

-0.65

utions

-0.65

itcher

-0.65

isons

-0.64

POSITIVE LOGITS

'd

0.81

 shouldn

0.81

 could

0.81

've

0.80

'll

0.79

 might

0.78

 cannot

0.77

 should

0.76

 ought

0.76

 conflic

0.74

Activations Density 0.305%

pronouns referring to the self or specific individuals

pronouns indicating the speaker or listener

No Comments

No Known Activations