INDEX

Explanations

abbreviations or acronyms related to geographical locations or organizations, particularly those with 'FL' in them

oai_token-act-pair · gpt-3.5-turbo

mentions of Florida

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

thodox

-0.74

uyomi

-0.73

nor

-0.72

ãĥīãĥ©ãĤ´ãĥ³

-0.72

 familiar

-0.72

selves

-0.69

 ç¥ŀ

-0.66

 Prayer

-0.65

 Hiroshima

-0.63

ItemTracker

-0.63

POSITIVE LOGITS

FL

1.10

oyd

0.98

ESH

0.98

UX

0.95

owship

0.94

OW

0.90

YING

0.90

acies

0.88

atical

0.88

TF

0.86

Activations Density 0.005%

abbreviations or acronyms related to geographical locations or organizations, particularly those with 'FL' in them

mentions of Florida

No Comments

No Known Activations