INDEX

Explanations

medical terms and references with numbers in them

oai_token-act-pair · gpt-3.5-turbo

references to academic publications or submissions

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

fits

-0.78

bil

-0.77

gress

-0.76

opian

-0.75

abal

-0.75

uten

-0.74

fit

-0.71

ALD

-0.70

hend

-0.68

 loudspe

-0.67

POSITIVE LOGITS

NetMessage

0.77

ãĥ¼ãĥ³

0.70

romeda

0.65

ļéĨĴ

0.65

THING

0.64

ãĥł

0.61

channelAvailability

0.58

Untitled

0.58

Īè

0.58

MFT

0.58

Activations Density 0.053%

medical terms and references with numbers in them

references to academic publications or submissions

No Comments

No Known Activations