INDEX

Explanations

phrases related to depth or in-depth analysis

oai_token-act-pair · gpt-3.5-turbo

references to "deeper" insights or understandings

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

advertising

-0.77

ery

-0.74

EED

-0.73

guard

-0.73

 WATCHED

-0.73

gmail

-0.71

eting

-0.71

Guard

-0.69

Counter

-0.68

 overrun

-0.67

POSITIVE LOGITS

 depths

0.97

 than

0.94

 depth

0.91

 insight

0.90

 layers

0.88

 deeper

0.84

 Than

0.83

 insights

0.82

dug

0.81

 vein

0.80

Activations Density 0.009%

phrases related to depth or in-depth analysis

references to "deeper" insights or understandings

No Comments

No Known Activations