INDEX

Explanations

instances indicating a focus on political or geopolitical developments

oai_token-act-pair · gpt-3.5-turbo

phrases related to discrepancies in data

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

ļéĨĴ

-0.77

OME

-0.75

 Advis

-0.71

Thumbnail

-0.69

ATIONS

-0.67

ģ«

-0.66

 Ethnic

-0.66

ATING

-0.65

İĭ

-0.65

ģĸ

-0.64

POSITIVE LOGITS

 Scotch

0.69

putable

0.66

dylib

0.66

riz

0.63

cipl

0.61

 Valiant

0.60

 cutter

0.59

ocious

0.59

ril

0.58

laden

0.56

Activations Density 0.000%

instances indicating a focus on political or geopolitical developments

phrases related to discrepancies in data

No Comments

No Known Activations