INDEX

Explanations

words or phrases that are closely related or have a strong connection to another word or concept

oai_token-act-pair · gpt-3.5-turbo

terms related to synonymy and equivalence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

annis

-0.71

vae

-0.68

 Bots

-0.68

uld

-0.67

GER

-0.67

amina

-0.62

Doctors

-0.61

UGE

-0.61

essee

-0.59

GRE

-0.59

POSITIVE LOGITS

 synonymous

1.35

entimes

0.79

ansk

0.77

inational

0.76

 Nadu

0.75

 symbol

0.73

icative

0.70

omorph

0.70

ophon

0.70

é¾įå¥ĳå£«

0.69

Activations Density 0.010%

words or phrases that are closely related or have a strong connection to another word or concept

terms related to synonymy and equivalence

No Comments

No Known Activations