INDEX

Explanations

studies or researches that show connections or correlations

oai_token-act-pair · gpt-3.5-turbo

references to research findings and studies

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Pledge

-0.74

 inflic

-0.73

 Shutdown

-0.66

 coward

-0.62

Thank

-0.62

Isn

-0.59

Interested

-0.59

 Rite

-0.58

 resign

-0.58

 Sending

-0.57

POSITIVE LOGITS

 shown

1.58

 demonstrated

1.19

 proven

1.13

 suggested

1.11

 been

1.08

 documented

1.08

 convinc

1.05

shown

1.04

 indicated

1.03

 revealed

1.03

Activations Density 0.130%

studies or researches that show connections or correlations

references to research findings and studies

No Comments

No Known Activations