INDEX

Explanations

negative statements or contradictions, where the latter part of the sentence contradicts the earlier part

oai_token-act-pair · gpt-3.5-turbo

negations and expressions of impossibility or inadequacy

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

only

-0.89

 merely

-0.82

 doubtless

-0.73

 PLUS

-0.66

 chiefly

-0.66

 unintentionally

-0.65

rimination

-0.64

 alternatively

-0.64

 inadvertently

-0.64

 falsely

-0.63

POSITIVE LOGITS

 darn

0.77

 fuckin

0.76

wow

0.74

enough

0.70

iability

0.67

hin

0.67

 enough

0.66

íķ

0.65

Õ

0.64

 fucking

0.63

Activations Density 0.068%

negative statements or contradictions, where the latter part of the sentence contradicts the earlier part

negations and expressions of impossibility or inadequacy

No Comments

No Known Activations