INDEX

Explanations

first-person pronouns

oai_token-act-pair · gpt-3.5-turbo

pronouns, particularly "I" and "we"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 monopoly

-0.86

 Speedway

-0.72

 torpedo

-0.71

 Nelson

-0.70

JFK

-0.65

 Forth

-0.65

NEO

-0.65

 Jarvis

-0.65

 impaired

-0.65

 Chop

-0.64

POSITIVE LOGITS

âĢ

1.61

ÃĥÃĤ

1.15

ÃĥÃĤÃĥÃĤ

1.10

Â´

1.06

Â«

1.05

Â»

1.03

Í

1.02

cffffcc

1.00

à¼

1.00

ÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤ

1.00

Activations Density 0.259%

first-person pronouns

pronouns, particularly "I" and "we"

No Comments

No Known Activations