INDEX

Explanations

phrases related to negative attributes or impacts

oai_token-act-pair · gpt-3.5-turbo

mentions of negative concepts or experiences

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

DOM

-0.87

conservancy

-0.81

ITNESS

-0.81

heet

-0.78

dropping

-0.77

hower

-0.77

plain

-0.77

pread

-0.77

abiding

-0.76

raltar

-0.76

POSITIVE LOGITS

 reinforcement

1.05

 spiral

0.90

 Negative

0.89

 gearing

0.86

 impact

0.85

 effects

0.85

 consequence

0.84

 feedback

0.82

 consequences

0.82

 publicity

0.82

Activations Density 0.023%

phrases related to negative attributes or impacts

mentions of negative concepts or experiences

No Comments

No Known Activations