INDEX

Explanations

phrases containing the name "Eddy" followed by a numerical value that is most likely a rating

oai_token-act-pair · gpt-3.5-turbo

words related to death or dying

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 prescribing

-0.61

 guides

-0.61

 drilling

-0.60

 Islands

-0.59

 Guide

-0.58

 bargaining

-0.57

 guide

-0.57

 poss

-0.56

 reserves

-0.56

ib

-0.56

POSITIVE LOGITS

dy

4.57

dies

1.73

die

1.51

dx

1.50

ds

1.36

di

1.35

din

1.34

der

1.32

Dy

1.28

dy

1.27

Activations Density 0.017%

phrases containing the name "Eddy" followed by a numerical value that is most likely a rating

words related to death or dying

No Comments

No Known Activations