INDEX

Explanations

mentions of specific job-related terms or positions

oai_token-act-pair · gpt-3.5-turbo

references to personnel in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

bow

-0.85

bows

-0.84

vid

-0.80

forth

-0.77

rous

-0.72

Brow

-0.72

fin

-0.70

ibia

-0.69

wit

-0.68

bolt

-0.68

POSITIVE LOGITS

 personnel

1.00

ilitary

0.90

 stationed

0.88

emetery

0.85

asures

0.84

 restraints

0.82

geant

0.82

andowski

0.81

 carriers

0.81

irtual

0.81

Activations Density 0.007%

mentions of specific job-related terms or positions

references to personnel in various contexts

No Comments

No Known Activations