INDEX

Explanations

names containing the word 'Bat'

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "Bat."

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

perse

-0.77

mble

-0.76

zanne

-0.70

ãĤĭ

-0.70

xia

-0.70

jri

-0.69

pressed

-0.66

essee

-0.65

ä

-0.65

×Ļ×

-0.64

POSITIVE LOGITS

room

0.96

tered

0.92

rooms

0.90

bat

0.89

oche

0.85

allion

0.84

illus

0.83

anas

0.82

avia

0.82

ting

0.81

Activations Density 0.010%

names containing the word 'Bat'

mentions of the word "Bat."

No Comments

No Known Activations