INDEX

Explanations

phrases related to opinions or arguments

oai_token-act-pair · gpt-3.5-turbo

expressions of uncertainty or ambiguity

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

 whilst

-0.73

 fucking

-0.70

 haha

-0.67

tv

-0.66

 ðŁĻĤ

-0.66

 fuckin

-0.64

âĢ¦âĢ¦

-0.62

Firstly

-0.62

)</

-0.62

 ðŁĺ

-0.60

POSITIVE LOGITS

 embold

0.71

 skeptics

0.70

 advocates

0.70

 critics

0.70

 skepticism

0.69

 lawmakers

0.68

 policymakers

0.68

 POLITICO

0.65

 resurg

0.65

 environmentalists

0.65

Activations Density 2.370%

phrases related to opinions or arguments

expressions of uncertainty or ambiguity

No Comments

No Known Activations