INDEX

Explanations

words related to health and medical procedures

oai_token-act-pair · gpt-3.5-turbo

special characters or symbols, specifically a unique character

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

USH

-0.71

ONG

-0.70

INTON

-0.69

ITS

-0.67

 sidx

-0.67

Reason

-0.65

uese

-0.64

 Failed

-0.64

 Goal

-0.64

 Scot

-0.63

POSITIVE LOGITS

zanne

0.67

clerosis

0.66

ija

0.66

eli

0.65

gettable

0.65

ryu

0.63

 contrace

0.63

furt

0.63

vel

0.63

iques

0.62

Activations Density 0.000%

words related to health and medical procedures

special characters or symbols, specifically a unique character

No Comments

No Known Activations