INDEX

Explanations

names of individuals and locations

oai_token-act-pair · gpt-3.5-turbo

names of individuals and keywords related to a specific study or drug

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

 Roots

-0.83

 salsa

-0.79

rock

-0.69

 Cash

-0.69

Jerry

-0.68

 Manhattan

-0.66

 Ginger

-0.66

 Tara

-0.66

phone

-0.66

roots

-0.65

POSITIVE LOGITS

Duc

2.66

 Moff

1.90

 unmanned

1.59

acan

1.42

 Fletcher

1.34

 Cout

1.15

Chu

1.06

 Admir

1.02

 Hutchinson

1.01

 Welch

0.98

Activations Density 0.060%

names of individuals and locations

names of individuals and keywords related to a specific study or drug

No Comments

No Known Activations