INDEX

Explanations

comparisons using the word "like"

oai_token-act-pair · gpt-3.5-turbo

expressions that compare or liken one thing to another

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

irements

-0.85

iership

-0.83

idates

-0.71

folios

-0.71

apers

-0.69

ugal

-0.67

ourse

-0.67

iets

-0.67

icipated

-0.67

enary

-0.66

POSITIVE LOGITS

lihood

1.09

 kidding

0.79

lier

0.76

 wink

0.72

wow

0.70

ably

0.70

liest

0.70

 nights

0.67

 crazy

0.67

 happens

0.65

Activations Density 0.097%

comparisons using the word "like"

expressions that compare or liken one thing to another

No Comments

No Known Activations