INDEX

Explanations

uncommon or specialized terms and technical language

oai_token-act-pair · gpt-3.5-turbo

expressions of uniqueness and individual qualities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ola

-0.91

WA

-0.89

zza

-0.82

Runner

-0.80

eeper

-0.78

 Mania

-0.76

olan

-0.75

Wilson

-0.73

ebus

-0.73

mia

-0.73

POSITIVE LOGITS

 instr

0.81

alt

0.78

ser

0.78

ãĥª

0.78

 securities

0.73

hig

0.73

ent

0.71

 rapt

0.71

 prote

0.70

ãĤ¤

0.69

Activations Density 0.151%

uncommon or specialized terms and technical language

expressions of uniqueness and individual qualities

No Comments

No Known Activations