INDEX

Explanations

terms related to approval, endorsement, and authority

oai_token-act-pair · gpt-3.5-turbo

words that are associated with food or dining

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

warts

-0.69

 Diesel

-0.68

 Corm

-0.61

Reloaded

-0.60

 Bulgarian

-0.58

 Penny

-0.58

 knot

-0.57

 unanswered

-0.57

 Norwegian

-0.57

ï¸ı

-0.57

POSITIVE LOGITS

illin

0.93

hof

0.85

asin

0.80

ments

0.79

isode

0.79

olitics

0.78

orthy

0.77

ciation

0.77

itures

0.76

itions

0.76

Activations Density 0.112%

terms related to approval, endorsement, and authority

words that are associated with food or dining

No Comments

No Known Activations