INDEX

Explanations

terms related to historical events and figures

oai_token-act-pair · gpt-3.5-turbo

statements that indicate the existence or attribution of something, especially involving subjective evaluations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Manifest

-0.63

 intimid

-0.60

 Labs

-0.59

 launches

-0.57

 interven

-0.57

 Citiz

-0.57

 Written

-0.56

 rains

-0.56

 hires

-0.56

 Rankings

-0.55

POSITIVE LOGITS

now

1.39

 hindsight

1.17

 remembered

1.16

now

1.06

 nowadays

0.98

 still

0.98

 credited

0.97

 recollection

0.97

 presently

0.96

 remember

0.93

Activations Density 0.551%

terms related to historical events and figures

statements that indicate the existence or attribution of something, especially involving subjective evaluations

No Comments

No Known Activations