INDEX

Explanations

words related to negative emotions or situations, particularly those involving distress

oai_token-act-pair · gpt-3.5-turbo

words related to distress or discomfort

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

 Knights

-0.66

 Leaf

-0.65

Gob

-0.64

Bec

-0.64

 elig

-0.63

 Become

-0.59

 Legion

-0.58

 Emerald

-0.57

 maker

-0.57

 Oath

-0.56

POSITIVE LOGITS

ressing

4.35

ressed

2.47

ress

2.08

ression

2.02

resses

1.88

ressive

1.71

ressor

1.56

orting

1.21

acting

1.05

 pressing

1.04

Activations Density 0.009%

words related to negative emotions or situations, particularly those involving distress

words related to distress or discomfort

No Comments

No Known Activations