INDEX

Explanations

mentions of detectives in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to detectives and investigative themes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ardless

-0.72

Fed

-0.70

ndra

-0.68

Obama

-0.68

eworld

-0.68

ende

-0.67

ignty

-0.67

aments

-0.66

rates

-0.64

uled

-0.64

POSITIVE LOGITS

 detective

3.57

 detectives

3.13

 Detective

2.78

 Detect

1.91

Detect

1.86

 investigator

1.79

 Sherlock

1.70

 Investigator

1.50

 investigators

1.46

 forensic

1.41

Activations Density 0.031%

mentions of detectives in various contexts

references to detectives and investigative themes

No Comments

No Known Activations