INDEX

Explanations

words related to denial and rebuttal in the context of accusations or claims

oai_token-act-pair · gpt-3.5-turbo

statements and claims related to denial or refutation

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tnc

-0.82

STEM

-0.70

 mosqu

-0.70

oscope

-0.69

odes

-0.68

rafted

-0.67

 Instrument

-0.66

glomer

-0.65

 dise

-0.65

Sensor

-0.64

POSITIVE LOGITS

 disav

1.43

 apologized

1.42

 denies

1.41

 refuted

1.40

 vehemently

1.40

 denied

1.34

 clarified

1.30

 apologised

1.28

 spokes

1.25

 apolog

1.24

Activations Density 0.527%

words related to denial and rebuttal in the context of accusations or claims

statements and claims related to denial or refutation

No Comments

No Known Activations