INDEX

Explanations

comparative phrases indicating a multiple increase or decrease

oai_token-act-pair · gpt-3.5-turbo

quantitative comparisons involving the word "times."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Reviewer

-1.05

aceous

-0.84

ogy

-0.78

CHAT

-0.78

arty

-0.77

irt

-0.76

liction

-0.73

OST

-0.70

XT

-0.69

RAW

-0.69

POSITIVE LOGITS

cale

0.98

 consecut

0.83

paces

0.75

 gestation

0.75

pan

0.72

cens

0.69

bp

0.69

hops

0.67

poons

0.66

cus

0.64

Activations Density 0.023%

comparative phrases indicating a multiple increase or decrease

quantitative comparisons involving the word "times."

No Comments

No Known Activations