INDEX

Explanations

phrases related to alternative names or aliases

oai_token-act-pair · gpt-3.5-turbo

the "aka" notation indicating alternative names or aliases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

eder

-0.72

ergic

-0.67

rehens

-0.65

egg

-0.63

urrent

-0.62

ochond

-0.61

 orbs

-0.61

Ö¼

-0.61

iners

-0.59

erest

-0.59

POSITIVE LOGITS

aka

0.96

hey

0.81

NULL

0.76

romeda

0.73

0000000000000000

0.71

 adolesc

0.70

roxy

0.69

CVE

0.68

nick

0.67

 alias

0.67

Activations Density 0.009%

phrases related to alternative names or aliases

the "aka" notation indicating alternative names or aliases

No Comments

No Known Activations