INDEX

Explanations

words related to military or security checkpoints

oai_token-act-pair · gpt-3.5-turbo

references to checkpoints

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

soever

-0.88

nder

-0.77

":"/

-0.76

mort

-0.75

morph

-0.74

uscript

-0.72

ankind

-0.71

tymology

-0.71

mol

-0.71

nian

-0.70

POSITIVE LOGITS

 checkpoints

1.24

 checkpoint

1.02

 clearance

0.90

 crossings

0.81

 Tanz

0.76

 kios

0.74

 divid

0.74

 Patrol

0.73

 plaza

0.73

 guards

0.71

Activations Density 0.016%

words related to military or security checkpoints

references to checkpoints

No Comments

No Known Activations