INDEX

Explanations

terms related to government officials, particularly those with specific names like "James" or "Jim"

oai_token-act-pair · gpt-3.5-turbo

mentions of individuals associated with government or intelligence positions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ahime

-0.74

opian

-0.69

[|

-0.68

xual

-0.68

Interstitial

-0.68

 Twins

-0.65

winner

-0.62

 Jagu

-0.61

 Nuggets

-0.61

 climax

-0.60

POSITIVE LOGITS

 Mattis

0.87

 Spicer

0.75

 Flynn

0.75

 Tillerson

0.74

 Comey

0.74

ghazi

0.74

reet

0.72

 Merrill

0.72

 Fallon

0.71

 briefed

0.71

Activations Density 0.188%

terms related to government officials, particularly those with specific names like "James" or "Jim"

mentions of individuals associated with government or intelligence positions

No Comments

No Known Activations