INDEX

Explanations

words related to being influenced negatively or harmed

oai_token-act-pair · gpt-3.5-turbo

references to individuals or groups that are impacted by a specific event or situation

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

WithNo

-0.73

OVA

-0.72

YC

-0.69

çĦ

-0.68

outed

-0.66

ALK

-0.66

outh

-0.66

ilts

-0.65

Vi

-0.64

otten

-0.63

POSITIVE LOGITS

 affected

1.03

 adversely

1.03

 effected

0.93

 impacted

0.89

 havoc

0.85

affected

0.79

 unaffected

0.77

 acutely

0.77

ados

0.75

 Affect

0.73

Activations Density 0.020%

words related to being influenced negatively or harmed

references to individuals or groups that are impacted by a specific event or situation

No Comments

No Known Activations