INDEX

Explanations

fragments of words with 'frag' in them

oai_token-act-pair · gpt-3.5-turbo

instances of the term "frag" and its variations, indicating a focus on fragmented concepts

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

eport

-0.82

istrate

-0.80

 constitu

-0.77

Reviewer

-0.74

busters

-0.73

eson

-0.71

yssey

-0.70

dayName

-0.69

hips

-0.69

ixed

-0.69

POSITIVE LOGITS

agne

0.79

ãĥĬ

0.76

izoph

0.75

ances

0.74

ANCE

0.74

ãĤ£

0.73

mental

0.69

ãĥĨãĤ£

0.68

itars

0.68

 Lank

0.67

Activations Density 0.039%

fragments of words with 'frag' in them

instances of the term "frag" and its variations, indicating a focus on fragmented concepts

No Comments

No Known Activations