INDEX

Explanations

numeric values and percentages

oai_token-act-pair · gpt-3.5-turbo

concepts related to division or disparity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Â»

-0.59

Rin

-0.59

 Donation

-0.55

 Shutterstock

-0.54

quote

-0.51

 Replay

-0.50

 Yugoslavia

-0.50

 Disclosure

-0.50

 Turks

-0.50

 Bohem

-0.49

POSITIVE LOGITS

abwe

0.73

dinand

0.70

zhen

0.69

thood

0.67

aurus

0.67

nesday

0.66

ween

0.65

anyl

0.65

oliath

0.64

ļé

0.64

Activations Density 0.763%

numeric values and percentages

concepts related to division or disparity

No Comments

No Known Activations