INDEX

Explanations

quantitative comparisons using the word "times"

oai_token-act-pair · gpt-3.5-turbo

phrases indicating multiplicative comparisons or ratios

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

cember

-0.78

services

-0.69

################################

-0.64

rals

-0.64

iku

-0.62

apolis

-0.61

LCS

-0.60

uctions

-0.60

liction

-0.59

bies

-0.59

POSITIVE LOGITS

 slower

0.89

 stronger

0.88

 greater

0.88

 faster

0.87

 louder

0.84

avier

0.84

 cheaper

0.83

 hotter

0.83

 worse

0.81

 heavier

0.80

Activations Density 0.041%

quantitative comparisons using the word "times"

phrases indicating multiplicative comparisons or ratios

No Comments

No Known Activations