INDEX

Explanations

mentions of a specific individual or surname, likely related to news stories or events

oai_token-act-pair · gpt-3.5-turbo

variations of the suffix "iers" in various contexts

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

shaw

-0.87

OUT

-0.74

MENTS

-0.74

Luck

-0.72

erest

-0.67

orses

-0.67

izu

-0.66

orate

-0.66

TL

-0.64

ulator

-0.63

POSITIVE LOGITS

hip

1.02

hips

0.90

pread

0.83

DragonMagazine

0.81

opher

0.79

aurus

0.78

ystem

0.76

pace

0.73

ingen

0.71

peak

0.71

Activations Density 0.016%

mentions of a specific individual or surname, likely related to news stories or events

variations of the suffix "iers" in various contexts

No Comments

No Known Activations