INDEX

Explanations

phrases related to organizations or services, specifically highlighting the National Park Service and similar entities

oai_token-act-pair · gpt-3.5-turbo

references to governmental or organizational services

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ussen

-0.74

luaj

-0.66

starter

-0.66

rous

-0.65

åĮ

-0.63

cause

-0.63

urate

-0.62

sbm

-0.61

ãĥ£

-0.61

umin

-0.61

POSITIVE LOGITS

 Employees

1.17

 Provider

1.04

 acad

0.86

 Corps

0.83

 Dogs

0.83

 Agency

0.82

 Accountability

0.81

 Desk

0.79

 Agents

0.78

 Officers

0.77

Activations Density 0.026%

phrases related to organizations or services, specifically highlighting the National Park Service and similar entities

references to governmental or organizational services

No Comments

No Known Activations