INDEX

Explanations

phrases containing the word "who"

oai_token-act-pair · gpt-3.5-turbo

the word "who" in various contexts, indicating a focus on identifying people or subjects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Bound

-0.62

 Processing

-0.60

Dos

-0.60

Georg

-0.58

 Mald

-0.58

Pad

-0.57

 Mayo

-0.57

 Delicious

-0.57

âĶĢâĶĢâĶĢâĶĢ

-0.56

ãĥ¼ãĥ³

-0.56

POSITIVE LOGITS

soever

0.92

iage

0.77

ever

0.69

oping

0.69

usalem

0.64

chwitz

0.63

pher

0.63

ileaks

0.63

 accomp

0.62

abama

0.61

Activations Density 0.076%

phrases containing the word "who"

the word "who" in various contexts, indicating a focus on identifying people or subjects

No Comments

No Known Activations