INDEX

Explanations

keywords related to government positions and titles

oai_token-act-pair · gpt-3.5-turbo

instances of the word "op" and various punctuation and numerical patterns related to context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

hots

-0.78

ococ

-0.75

otes

-0.75

-0.73

ãĥĻ

-0.73

-0.70

-0.69

-0.67

 Reese

-0.66

 Scarlet

-0.65

POSITIVE LOGITS

1.26

dor

1.19

1.18

dl

1.12

dt

1.10

Ds

1.09

ds

1.06

 Dixon

1.03

DI

1.03

dh

1.02

Activations Density 0.265%

keywords related to government positions and titles

instances of the word "op" and various punctuation and numerical patterns related to context

No Comments

No Known Activations