INDEX

Explanations

proper nouns, specifically names like 'Jin', 'Rune', 'Nex', and 'Unreal'

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to characters and games

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

@@@@@@@@

-0.82

eed

-0.77

IRT

-0.73

ARDS

-0.72

 horm

-0.71

ARD

-0.69

Lean

-0.68

\\\\

-0.67

\/\/

-0.67

chairs

-0.66

POSITIVE LOGITS

jin

1.15

ominated

0.90

onymous

0.85

ovych

0.81

etheus

0.79

opolis

0.79

furt

0.79

ameless

0.78

vernment

0.78

okin

0.78

Activations Density 0.031%

proper nouns, specifically names like 'Jin', 'Rune', 'Nex', and 'Unreal'

proper nouns related to characters and games

No Comments

No Known Activations