INDEX

Explanations

words related to theaters or performance venues

oai_token-act-pair · gpt-3.5-turbo

mentions of various theaters

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

 predetermined

-0.74

 strokes

-0.72

 sheet

-0.71

alysed

-0.70

 trouble

-0.70

 sheets

-0.69

ipeg

-0.67

 tide

-0.66

vernment

-0.65

assed

-0.65

POSITIVE LOGITS

 District

1.17

 Parkway

1.09

 Plaza

1.09

 Association

1.06

 Hotel

1.05

 Authority

1.04

 Orchestra

1.01

 Improvement

1.00

 Trust

0.98

 Company

0.98

Activations Density 0.138%

words related to theaters or performance venues

mentions of various theaters

No Comments

No Known Activations