INDEX

Explanations

specific proper nouns, particularly the abbreviated name "Ole" from "Ole Miss"

oai_token-act-pair · gpt-4-turbo

mentions of "Ole Miss" and its associated context or events

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

mentions of the term "Ole Miss"

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ãĤ½

-0.78

ULT

-0.75

ãĤ¢ãĥ«

-0.73

travel

-0.70

Phoenix

-0.70

DERR

-0.67

stroke

-0.67

ãĥĩ

-0.66

ãĥŁ

-0.65

addock

-0.64

POSITIVE LOGITS

Ole

1.09

phan

0.89

terness

0.84

oshenko

0.80

enko

0.80

 Klux

0.74

etsk

0.74

oko

0.74

yss

0.74

cular

0.74

Activations Density 0.007%

specific proper nouns, particularly the abbreviated name "Ole" from "Ole Miss"

mentions of "Ole Miss" and its associated context or events

mentions of the term "Ole Miss"

No Comments

No Known Activations