INDEX

Explanations

proper nouns and titles of individuals

oai_token-act-pair · gpt-3.5-turbo

phrases related to political figures and their actions or attributes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

odge

-0.67

irie

-0.67

aways

-0.66

 necessities

-0.66

 bedrooms

-0.66

EMS

-0.63

 cohesion

-0.63

 tents

-0.62

urry

-0.62

 preliminary

-0.61

POSITIVE LOGITS

who

1.61

 whom

1.52

whose

1.43

who

1.42

 whose

1.27

 himself

1.22

 Himself

1.19

someone

1.15

 extraord

1.07

aka

1.07

Activations Density 0.567%

proper nouns and titles of individuals

phrases related to political figures and their actions or attributes

No Comments

No Known Activations