INDEX

Explanations

personal pronouns and related verbs indicating hypothetical scenarios or decision-making

oai_token-act-pair · gpt-3.5-turbo

repeated first-person pronouns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

quartered

-0.76

mud

-0.65

Ts

-0.65

IOR

-0.62

 Bearing

-0.62

bats

-0.61

MENTS

-0.60

interstitial

-0.60

folios

-0.59

 Emanuel

-0.58

POSITIVE LOGITS

 succeed

0.86

 weren

0.83

're

0.77

 happen

0.76

 were

0.75

 someday

0.73

 grap

0.72

 fail

0.72

lisher

0.69

 wanna

0.69

Activations Density 0.145%

personal pronouns and related verbs indicating hypothetical scenarios or decision-making

repeated first-person pronouns

No Comments

No Known Activations