INDEX

Explanations

names or terms related to various locations

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names of individuals and places, as well as references to entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

odium

-0.77

AMS

-0.67

SAP

-0.66

GREEN

-0.63

 rack

-0.61

iox

-0.61

 Logged

-0.61

 barr

-0.61

adel

-0.60

 intermitt

-0.59

POSITIVE LOGITS

kun

1.98

arest

1.71

 Claus

1.65

 Pluto

1.53

 Satan

1.37

 Dracula

1.25

 Psycho

1.18

 psycho

1.18

anova

1.16

 caus

1.16

Activations Density 0.044%

names or terms related to various locations

proper nouns, specifically names of individuals and places, as well as references to entities

No Comments

No Known Activations