INDEX

Explanations

names of individuals with the word "gin" in them

oai_token-act-pair · gpt-3.5-turbo

repeated occurrences of the term "gin" and names related to medical or toxicological contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Breach

-0.81

awei

-0.70

 Immunity

-0.67

 Parables

-0.67

 EDITION

-0.66

 McDonnell

-0.65

 Settlement

-0.64

 Cutter

-0.63

 Humanity

-0.62

ship

-0.62

POSITIVE LOGITS

uine

1.14

gin

1.12

ppo

0.95

etimes

0.91

xon

0.90

oooooooooooooooo

0.82

 acids

0.81

orgetown

0.79

plates

0.79

xes

0.79

Activations Density 0.033%

names of individuals with the word "gin" in them

repeated occurrences of the term "gin" and names related to medical or toxicological contexts

No Comments

No Known Activations