INDEX

Explanations

words related to people with the prefix "Di"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "Di" followed by various identifiers or titles related to individuals or organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

manship

-1.00

sburgh

-0.86

éĹĺ

-0.86

INGS

-0.79

 Twain

-0.76

nesday

-0.74

ilater

-0.73

ãģĨ

-0.73

ãĤ¤ãĥĪ

-0.73

CLASSIFIED

-0.72

POSITIVE LOGITS

abol

1.26

abolic

1.19

agram

1.07

urnal

1.06

abetic

1.05

ocese

1.02

ploma

1.02

agon

1.01

apers

1.00

abetes

0.98

Activations Density 0.008%

words related to people with the prefix "Di"

occurrences of the term "Di" followed by various identifiers or titles related to individuals or organizations

No Comments

No Known Activations