INDEX

Explanations

names of hotels or event venues

oai_token-act-pair · gpt-3.5-turbo

references to hotels and theaters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

theless

-0.72

 strokes

-0.71

 sheets

-0.69

etheless

-0.68

 knight

-0.67

copy

-0.66

 sheet

-0.66

genre

-0.65

gpu

-0.64

 favour

-0.64

POSITIVE LOGITS

 Hotel

1.14

 District

1.13

 Plaza

1.10

 Collection

1.04

 Orchestra

1.03

 Parkway

1.02

 Trust

1.01

 Authority

1.00

 Room

1.00

 Improvement

0.99

Activations Density 0.124%

names of hotels or event venues

references to hotels and theaters

No Comments

No Known Activations