INDEX

Explanations

instances where investigations are being conducted

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "investigate" and its variations

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

knife

-0.78

aa

-0.76

VP

-0.72

dos

-0.71

perm

-0.67

opp

-0.66

Jet

-0.66

relative

-0.65

align

-0.65

joining

-0.65

POSITIVE LOGITS

 investigating

1.05

 investigates

0.98

 investigation

0.96

 investigated

0.90

 investigations

0.90

 investigate

0.87

çīĪ

0.84

 Investigation

0.82

 Investigative

0.82

ĸļ

0.81

Activations Density 0.015%

instances where investigations are being conducted

occurrences of the term "investigate" and its variations

No Comments

No Known Activations