INDEX

Explanations

comparative terms expressing improvement or advantage

oai_token-act-pair · gpt-3.5-turbo

phrases expressing improvement or increased quality

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

lees

-0.91

events

-0.91

coins

-0.84

 Dates

-0.79

hops

-0.78

ients

-0.78

casts

-0.78

ulations

-0.77

guards

-0.76

 Committees

-0.76

POSITIVE LOGITS

 version

1.15

 option

0.99

iator

0.97

 approximation

0.96

 explanation

0.96

than

0.95

 solution

0.92

 than

0.92

 approach

0.91

 role

0.90

Activations Density 0.119%

comparative terms expressing improvement or advantage

phrases expressing improvement or increased quality

No Comments

No Known Activations