INDEX

Explanations

comparative phrases indicating significance or preference

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the best or most significant options and advantages

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ipel

-0.79

avis

-0.67

azard

-0.66

renheit

-0.64

utics

-0.64

hak

-0.62

nothing

-0.62

zilla

-0.62

jong

-0.62

apesh

-0.62

POSITIVE LOGITS

 reason

1.15

 downside

1.10

 thing

1.09

 question

1.06

 caveat

1.04

 takeaway

1.04

 drawback

1.03

 distinguishing

1.02

 concern

0.97

 exception

0.95

Activations Density 0.228%

comparative phrases indicating significance or preference

phrases indicating the best or most significant options and advantages

No Comments

No Known Activations