INDEX

Explanations

information related to specific numerical values and events such as dates, percentages, locations, and names

oai_token-act-pair · gpt-3.5-turbo

numerical data and statistics

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

"))

-0.70

death

-0.70

iverse

-0.67

onz

-0.66

helm

-0.64

"]

-0.63

NK

-0.63

]"

-0.62

uts

-0.62

lass

-0.61

POSITIVE LOGITS

 topping

1.01

 being

1.01

 preferring

0.96

 intervening

0.89

 needing

0.88

 seeming

0.88

 reaching

0.88

 echoing

0.87

 boasting

0.87

 appearing

0.87

Activations Density 0.412%

information related to specific numerical values and events such as dates, percentages, locations, and names

numerical data and statistics

No Comments

No Known Activations