INDEX

Explanations

percentages expressed in parentheses

oai_token-act-pair · gpt-3.5-turbo

percentages expressed in parentheses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

zbollah

-0.71

uran

-0.70

 artif

-0.70

 tyrann

-0.67

 shack

-0.66

imus

-0.65

 princ

-0.65

mus

-0.64

oria

-0.64

Syd

-0.64

POSITIVE LOGITS

theless

0.87

ecided

0.78

osuke

0.78

erest

0.77

okers

0.75

 Decay

0.72

+)

0.67

 Unsure

0.67

TPS

0.66

apest

0.66

Activations Density 0.018%

percentages expressed in parentheses

percentages expressed in parentheses

No Comments

No Known Activations