INDEX

Explanations

It looks for references to past events or projects

oai_token-act-pair · gpt-3.5-turbo

sentences that express opinions or conclusions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 satell

-0.82

 torpedo

-0.80

 daring

-0.78

 mushroom

-0.77

quartered

-0.74

 fermented

-0.72

ascus

-0.72

 fancy

-0.71

 bindings

-0.71

 foam

-0.71

POSITIVE LOGITS

âĢ

2.10

 âĢ

1.90

 âĢº

1.53

Â»

1.42

 âĶĤ

1.39

 âĶ

1.35

âĢİ

1.33

 âĢİ

1.32

 âĸº

1.29

 âľ

1.26

Activations Density 0.377%

It looks for references to past events or projects

sentences that express opinions or conclusions

No Comments

No Known Activations