INDEX

Explanations

names or terms containing the word "Low" followed by a numerical rating

oai_token-act-pair · gpt-3.5-turbo

occurrences of the name "Low" or variations of it

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

UAL

-0.86

clave

-0.85

andise

-0.81

arium

-0.78

natureconservancy

-0.74

zeb

-0.69

ãĤ·ãĥ£

-0.68

andum

-0.66

ansas

-0.65

netic

-0.65

POSITIVE LOGITS

ered

1.02

enthal

0.93

rey

0.90

down

0.89

est

0.85

ering

0.78

brow

0.76

zzi

0.74

estone

0.73

reys

0.72

Activations Density 0.065%

names or terms containing the word "Low" followed by a numerical rating

occurrences of the name "Low" or variations of it

No Comments

No Known Activations