INDEX

Explanations

actions and events related to people being affected or involved in various incidents

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Locations

-0.90

iencies

-0.85

 Units

-0.84

 Zombies

-0.74

 Areas

-0.73

 engagements

-0.72

 Lines

-0.72

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.71

 Functions

-0.71

 Factors

-0.71

POSITIVE LOGITS

 stabbed

0.92

 accused

0.87

 named

0.85

 suing

0.82

sole

0.82

 murdered

0.81

 claiming

0.81

 accidentally

0.78

 kidnapped

0.78

 crowned

0.77

Activations Density 0.302%

actions and events related to people being affected or involved in various incidents

No Comments

No Known Activations