INDEX

Explanations

proper nouns that appear to be related to individuals in various contexts

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names associated with significant historical or cultural references

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

iture

-0.45

redients

-0.45

owship

-0.44

anwhile

-0.43

ATIONAL

-0.43

auga

-0.41

 sake

-0.40

selves

-0.40

iage

-0.39

 Nadu

-0.39

POSITIVE LOGITS

0.44

²

0.43

ł

0.43

lov

0.43

force

0.43

¾

0.42

ill

0.42

zar

0.41

open

0.41

½

0.40

Activations Density 0.679%

proper nouns that appear to be related to individuals in various contexts

proper nouns and names associated with significant historical or cultural references

No Comments

No Known Activations