INDEX

Explanations

the word "Nemesis" and related terms

oai_token-act-pair · gpt-3.5-turbo

references to specific names or terms related to "Nem."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-0.77

ãĥ³ãĤ¸

-0.71

ãĥ¼ãĥĨ

-0.70

FORM

-0.68

 USAF

-0.68

 Xander

-0.67

 Scots

-0.66

shapeshifter

-0.65

 Wasteland

-0.64

 Protestant

-0.64

POSITIVE LOGITS

eny

1.04

Nem

0.99

ecast

0.97

acles

0.95

omon

0.91

esis

0.88

daq

0.88

ilo

0.88

oji

0.87

enta

0.85

Activations Density 0.008%

the word "Nemesis" and related terms

references to specific names or terms related to "Nem."

No Comments

No Known Activations