INDEX

Explanations

phrases related to specific locations or events, particularly focusing on incidents and locations like San Bernardino

oai_token-act-pair · gpt-3.5-turbo

references to specific locations related to shootings

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ATER

-0.76

adj

-0.72

BY

-0.69

fare

-0.69

Ole

-0.67

WIND

-0.67

OPLE

-0.67

PLAY

-0.66

 Viking

-0.65

Minimum

-0.63

POSITIVE LOGITS

 Bernardino

1.44

icides

0.92

 Rica

0.88

 County

0.88

itized

0.86

iage

0.84

itiz

0.80

andon

0.77

 condem

0.76

inia

0.75

Activations Density 0.002%

phrases related to specific locations or events, particularly focusing on incidents and locations like San Bernardino

references to specific locations related to shootings

No Comments

No Known Activations