INDEX

Explanations

terms related to investigative reporting

oai_token-act-pair · gpt-3.5-turbo

terms related to investigative journalism

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

yip

-1.03

creen

-0.88

jri

-0.88

unin

-0.86

ogun

-0.83

olen

-0.81

hops

-0.79

generic

-0.78

ivari

-0.77

aus

-0.74

POSITIVE LOGITS

 Investigative

1.03

 investigative

0.98

 journalism

0.93

 investigations

0.93

 Investigations

0.88

 investigation

0.88

 probing

0.87

 journalist

0.85

 Investigation

0.82

 reporting

0.82

Activations Density 0.029%

terms related to investigative reporting

terms related to investigative journalism

No Comments

No Known Activations