INDEX

Explanations

questions posed rhetorically for confirmation

oai_token-act-pair · gpt-3.5-turbo

rhetorical questions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

apan

-0.72

rament

-0.68

shaw

-0.66

 foreground

-0.64

 binge

-0.62

apers

-0.61

 lobster

-0.60

thro

-0.60

 slam

-0.60

 background

-0.59

POSITIVE LOGITS

 Nope

0.96

 Wouldn

0.91

 Yeah

0.87

 Anyway

0.86

Why

0.84

 Especially

0.84

Isn

0.84

 Surely

0.83

 Alright

0.81

 Maybe

0.80

Activations Density 0.065%

questions posed rhetorically for confirmation

rhetorical questions

No Comments

No Known Activations