INDEX

Explanations

commands or requests for action

oai_token-act-pair · gpt-3.5-turbo

second-person pronouns and related phrases indicating conditional scenarios

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Eleven

-0.69

 Forth

-0.66

IB

-0.65

Reviewer

-0.64

 Resurrection

-0.63

 Twelve

-0.63

 Bones

-0.62

 Ammunition

-0.62

 Results

-0.61

 Invention

-0.60

POSITIVE LOGITS

're

0.94

eded

0.90

ever

0.84

 encount

0.84

 wished

0.79

'd

0.78

 weren

0.77

 knew

0.76

anium

0.76

've

0.75

Activations Density 0.148%

commands or requests for action

second-person pronouns and related phrases indicating conditional scenarios

No Comments

No Known Activations