INDEX

Explanations

words related to chemical compounds, particularly those ending with "iol"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "iol"

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

 heels

-0.70

 flagged

-0.67

AK

-0.63

 Rails

-0.61

TS

-0.60

Alt

-0.60

 fabrication

-0.60

 harassment

-0.59

 accusations

-0.59

 Point

-0.58

POSITIVE LOGITS

iol

4.70

iov

1.54

iola

1.52

io

1.39

iop

1.36

iology

1.32

ios

1.32

ione

1.25

iot

1.19

iological

1.19

Activations Density 0.012%

words related to chemical compounds, particularly those ending with "iol"

occurrences of the substring "iol"

No Comments

No Known Activations