INDEX

Explanations

named entities related to North Korea and health conditions

oai_token-act-pair · gpt-3.5-turbo

terms related to North Korea (NK) and its intelligence activities

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

izont

-0.76

 grounding

-0.73

 Grande

-0.70

faces

-0.68

 Element

-0.68

amera

-0.66

rha

-0.66

 Tinker

-0.65

veyard

-0.64

 flown

-0.62

POSITIVE LOGITS

VD

1.19

icts

0.81

nut

0.75

uint

0.73

apeake

0.72

hua

0.71

orea

0.70

enders

0.70

Kw

0.70

NK

0.69

Activations Density 0.020%

named entities related to North Korea and health conditions

terms related to North Korea (NK) and its intelligence activities

No Known Activations